The role of lumbar puncture (LP) in syphilis is controversial, particularly in HIV-infected patient. Treponema pallidum invades the central nervous system (CNS) early in disease, and benzathine penicillin G (BPG), the recommended treatment for uncomplicated syphilis, does not achieve treponemacidal CSF drug levels. Serologically defined treatment failure is common in HIV-infected syphilis patients, perhaps due to untreated neurosyphilis (NS). Patients at highest risk for NS can be identified based on serum Rapid Plasma Reagin (RPR) titer and peripheral blood CD4 count. While most experts agree that patients with syphilis who have neurological symptoms should undergo LP, the approach to patients without neurological complaints is less clear. Some experts argue that a strategy of immediate LP will avert neurological morbidity. Others argue that symptomatic NS is uncommon after BPG and LP can be reserved for those who fail treatment. In this proposal, we will test the hypothesis that a strategy of immediate LP i high-risk individuals, followed by therapy based on cerebrospinal fluid (CSF) examination, results in better serological outcomes in HIV-infected syphilis patients. A second area of uncertainty is whether neuroinvasion impacts cognition in individuals who might otherwise be considered neurologically asymptomatic. We show that HIV-infected patients with previous syphilis, especially those with CSF abnormalities consistent with NS, are cognitively impaired. We will test the hypothesis that CNS involvement by T. pallidum has cognitive and functional consequences, and that a strategy of immediate LP in high-risk individuals, followed by therapy based CSF examination, results in better cognitive and functional outcomes in HIV-infected syphilis patients. Our research plan takes advantage of established expertise and infrastructure for study of syphilis, NS and cognitive function in HIV-infected patients. We will randomize HIV-infected patients at high risk for NS based on serum RPR e1:32 or CD4 d350/ul into two groups: LP (treatment guided by CSF evaluation: Confirmed NS [treated for NS] or No NS [treated with BPG]) and No LP (treatment not guided by CSF evaluation: all receive BPG). The Specific Aims are: 1) Determine if a strategy of immediate LP results in better serological outcomes in HIV-infected patients with syphilis. Hypothesis: the No LP group, which includes undiagnosed and untreated NS patients, will have a poorer serological treatment response than the LP group; 2) Identify cognitive and functional impairment in HIV-infected patients at diagnosis of syphilis. Hypothesis: cognitive and functional performance will be poorer in the Confirmed NS patients compared to the No NS patients. Because the No LP group includes individuals with undiagnosed NS, they will perform more poorly than the No NS patients; 3) Determine if a strategy of immediate LP results in better cognitive and functional outcomes in HIV-infected patients with syphilis. Hypothesis: the No LP group, which includes undiagnosed and untreated NS patients, will have less improvement in cognition and function after therapy than the LP group.